# grid questions

 Register Blogs Members List Search Today's Posts Mark Forums Read

 November 27, 2007, 17:26 grid questions #1 toni Guest   Posts: n/a I'm using gambit to mesh and solving in fluent. I'm modeling downward flow in a pipe with elevated wall temperatures to study how buoyancy forces can affect the flow reversal and the behavior of this reversal. A few question: 1) Grid independence. I understand to find this you refine the mesh until the solution does not appreciably change. But what about different mesh volumes types: cooper, tet, etc. How do I know which type to use? My answer is different depending on tet or cooper. 2) Since the interesting part of this flow is near the wall, I tried two things: one was just a fine tet mesh. The other was to use boundary layers near the wall, then a cooper type for the interior. What is recommended for something like this? 3) what is meant by structured, unstructured mesh? Thanks toni

 November 27, 2007, 17:47 Re: grid questions #2 raj Guest   Posts: n/a Structured is like using cooper element and unstructed is like using tet element...for most cases try to use structeured mesh..but most practical problem it's not possible..so wherever possible use structured (by domain decomposition method) and other places use unstructured... for ur cases to capture wall stress effects, use BL at one face and cooper to next face..i.e first do plane mesh ( 2D mesh in one face) and extrude to next face... Mesh independency is first say ur using 10000 elements and ur getting some pressure drop...if u increase ur elemets to 20000 then ur pressure drop (or any other parameter of interst) varies significantly then ur second results is correct but not well enough again do ur analysis for higher elements check the results...if ur values not changing much say values are <5% then u can stop ur analysis and use ur results for further calculation... But the above said things cant be done for large practical problem where complicated Geometry and size of problem is more involved...so that cases Mesh Independency could be avaoided by using more element used during first time meshing with proper care at the wall's taken care...Again the results are may or may not be accurate...after all, these are numeruical technique, means u never achive 100% close results... Ur case try to use BL option at the wall's are have more elements at the wall..also take care of elemnt size variation..means change of the size of the elements should be gradual (like Aspect Ratio).... Gud Luck......

 December 3, 2007, 02:17 Re: grid questions #3 Ramesh. K Guest   Posts: n/a Hi, Structured mesh means if u know the information of the adjacent cells information and the distance between the cells then that type of mesh is called structured mesh its like "who is whose neighbor and at what distance they are" you need to store this information explicitly in ur data structure for grid generation. but in the case of unstructured grid u need not store this information. we use this information while solving. if ur solving viscous flows always fine mesh like the one in boundayr layers in gambit are preferable to the tet or triangulation at the boundary. because at the boundary the derivatives are in perpendicular direction to the wall so cartesian mesh with fine mesh at the boundary is suitable.

 December 5, 2007, 14:32 Re: grid questions #4 toni Guest   Posts: n/a thank you both for your recommendations. toni

 Thread Tools Display Modes Linear Mode

 Posting Rules You may not post new threads You may not post replies You may not post attachments You may not edit your posts BB code is On Smilies are On [IMG] code is On HTML code is OffTrackbacks are On Pingbacks are On Refbacks are On Forum Rules

 Similar Threads Thread Thread Starter Forum Replies Last Post albcem OpenFOAM 0 May 5, 2009 14:17 Sas Main CFD Forum 8 September 2, 2005 08:36 leo Siemens 3 April 9, 2003 01:35 Hans Klaufus CFX 1 June 28, 2000 16:43 Hansong Hang Main CFD Forum 12 September 16, 1998 22:26

All times are GMT -4. The time now is 02:59.